{
  "id": 63709,
  "title": "MLflow + Hyperopt workflow",
  "url": "/competitions/airbus-ship-detection/discussion/63709",
  "author_name": "Yassine Alouini",
  "post_date": "2018-08-20T13:59:36.182000",
  "votes": 4,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Hello everyone. </p>\n\n<p>I am releasing a kernel that details a hyperparameters optimisation workflow. It uses hyperopt for the optimization part and mlflow for tracking the various runs (a run is defined by its hyperparameters).</p>\n\n<p>Notice that this is a WIP and a lot of things could (and will) be improved. </p>\n\n<p>Link to the kernel where you will find more details: <a href=\"https://www.kaggle.com/yassinealouini/mlflow-hyperopt-u-net-workflow\">https://www.kaggle.com/yassinealouini/mlflow-hyperopt-u-net-workflow</a>.</p>\n\n<p>Enjoy! </p>",
  "messages": [
    {
      "id": 372870,
      "postDate": "2018-08-20T13:59:36.183Z",
      "content": "<p>Hello everyone. </p>\n\n<p>I am releasing a kernel that details a hyperparameters optimisation workflow. It uses hyperopt for the optimization part and mlflow for tracking the various runs (a run is defined by its hyperparameters).</p>\n\n<p>Notice that this is a WIP and a lot of things could (and will) be improved. </p>\n\n<p>Link to the kernel where you will find more details: <a href=\"https://www.kaggle.com/yassinealouini/mlflow-hyperopt-u-net-workflow\">https://www.kaggle.com/yassinealouini/mlflow-hyperopt-u-net-workflow</a>.</p>\n\n<p>Enjoy! </p>",
      "rawMarkdown": "Hello everyone. \n\nI am releasing a kernel that details a hyperparameters optimisation workflow. It uses hyperopt for the optimization part and mlflow for tracking the various runs (a run is defined by its hyperparameters).\n\nNotice that this is a WIP and a lot of things could (and will) be improved. \n\nLink to the kernel where you will find more details: https://www.kaggle.com/yassinealouini/mlflow-hyperopt-u-net-workflow.\n\nEnjoy! ",
      "votes": 4
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "372870": "Hello everyone. \n\nI am releasing a kernel that details a hyperparameters optimisation workflow. It uses hyperopt for the optimization part and mlflow for tracking the various runs (a run is defined by its hyperparameters).\n\nNotice that this is a WIP and a lot of things could (and will) be improved. \n\nLink to the kernel where you will find more details: https://www.kaggle.com/yassinealouini/mlflow-hyperopt-u-net-workflow.\n\nEnjoy! "
  }
}